Re: [PATCH] usb: dwc3: usb/dwc3: fake dissconnect event when turn off pullup

[Felipe Balbi]

Hi,

changbin.du@xxxxxxxxx writes:
> From: "Du, Changbin" <changbin.du@xxxxxxxxx>
>
> The dwc3 controller can't generate a disconnect event after it is
> stopped. Thus gadget dissconnect callback is not invoked when do

not generating disconnect IRQ after you drop Run/Stop is expected.

> soft dissconnect. Call dissconnect here to workaround this issue.

I'm rather sure this is a bug elsewhere. How do you trigger this ?

> Note, most time we still see disconnect be called that because
> it is invoked by dwc3_gadget_reset_interrupt(). But if we
> disconnect cable once pullup disabled quickly, issue can be
> observed.

I can't understand what you're trying to say with this.

> Signed-off-by: Du, Changbin <changbin.du@xxxxxxxxx>
> ---
>  drivers/usb/dwc3/gadget.c | 33 ++++++++++++++++++++++++---------
>  1 file changed, 24 insertions(+), 9 deletions(-)
>
> diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c
> index 8e4a1b1..cd73187 100644
> --- a/drivers/usb/dwc3/gadget.c
> +++ b/drivers/usb/dwc3/gadget.c
> @@ -1566,6 +1566,15 @@ static int dwc3_gadget_run_stop(struct dwc3 *dwc, int is_on, int suspend)
>  	return 0;
>  }
>  
> +static void dwc3_disconnect_gadget(struct dwc3 *dwc)
> +{
> +	if (dwc->gadget_driver && dwc->gadget_driver->disconnect) {
> +		spin_unlock(&dwc->lock);
> +		dwc->gadget_driver->disconnect(&dwc->gadget);
> +		spin_lock(&dwc->lock);
> +	}
> +}
> +
>  static int dwc3_gadget_pullup(struct usb_gadget *g, int is_on)
>  {
>  	struct dwc3		*dwc = gadget_to_dwc(g);
> @@ -1575,6 +1584,21 @@ static int dwc3_gadget_pullup(struct usb_gadget *g, int is_on)
>  	is_on = !!is_on;
>  
>  	spin_lock_irqsave(&dwc->lock, flags);
> +	/**
> +	 * WORKAROUND: The dwc3 controller can't generate a disconnect
> +	 * event after it is stopped. Thus gadget dissconnect callback
> +	 * is not invoked when do soft dissconnect. Call dissconnect
> +	 * here to workaround this issue.
> +	 * Note, most time we still see disconnect be called that because
> +	 * it is invoked by dwc3_gadget_reset_interrupt(). But if we
> +	 * disconnect cable once pullup disabled quickly, issue can be
> +	 * observed.
> +	 */
> +	if (!is_on && (dwc->gadget.speed != USB_SPEED_UNKNOWN)) {
> +		dev_dbg(dwc->dev, "fake dissconnect event on pullup off\n");
> +		dwc3_disconnect_gadget(dwc);
> +		dwc->gadget.speed = USB_SPEED_UNKNOWN;
> +	}

this is *really* wrong. You shouldn't be calling pullup directly. This
patch looks wrong and you didn't even explain how to trigger this
problem; when does the problem happen ?

Gadget load/unload does the right thing here, so that can't be the
case. We also do the right thing on soft_connect sysfs file:

static ssize_t usb_udc_softconn_store(struct device *dev,
		struct device_attribute *attr, const char *buf, size_t n)
{
	struct usb_udc		*udc = container_of(dev, struct usb_udc, dev);

	if (!udc->driver) {
		dev_err(dev, "soft-connect without a gadget driver\n");
		return -EOPNOTSUPP;
	}

	if (sysfs_streq(buf, "connect")) {
		usb_gadget_udc_start(udc);
		usb_gadget_connect(udc->gadget);
	} else if (sysfs_streq(buf, "disconnect")) {
		usb_gadget_disconnect(udc->gadget);
		udc->driver->disconnect(udc->gadget);
		usb_gadget_udc_stop(udc);
	} else {
		dev_err(dev, "unsupported command '%s'\n", buf);
		return -EINVAL;
	}

	return n;
}
static DEVICE_ATTR(soft_connect, S_IWUSR, NULL, usb_udc_softconn_store);

So really, what _is_ the problem ?

-- 
balbi

Attachment: signature.asc
Description: PGP signature